Selected work

Representative Papers

Kruskal-EDS: Edge Dynamic Stratification

Mar 02, 2026

This work addresses the inefficiency of the classical Kruskal’s algorithm, whose Θ(m log m) time complexity—stemming from global edge sorting—becomes prohibitive on sparse graphs or those with heavy-tailed weight distributions. The authors propose a dynamic layered variant of Kruskal’s algorithm that circumvents full sorting through a three-stage pipeline: first, sampling to estimate the weight distribution; second, adaptively constructing a layered structure inspired by the Birkhoff ergodic theorem to dynamically determine the optimal number of layers k*; and third, processing edges layer-by-layer with early termination. By integrating statistical sampling, binary search over quantile-based boundaries, and per-layer sorting, the method achieves near-linear complexity on sparse or skewed-weight graphs. Experiments across 14 graph types confirm correctness in 12 cases, demonstrating up to a 10× speedup over standard Kruskal and a 33× reduction in sorting operations.

0 citationsRead paper

Formal Semantic Control over Language Models

Jan 31, 2026

Current language models often produce semantic representations that lack interpretability and controllability, hindering localized, quasi-symbolic, and compositional semantic manipulation. To address this limitation, this work proposes a novel approach within the variational autoencoder (VAE) framework that explicitly models the geometric structure of the latent space to achieve disentangled, isolated, and directionally controllable semantic features. By systematically enhancing the interpretability and structural organization of the latent space, the method enables precise and reliable semantic control in tasks such as sentence generation and explanatory natural language inference (Explanatory NLI). This advancement significantly improves the model’s capacity for high-level semantic manipulation while maintaining fidelity to the underlying linguistic structures.

0 citationsRead paper

VEIGAR: View-consistent Explicit Inpainting and Geometry Alignment for 3D object Removal

Jun 13, 2025

To address the limitations of existing 3D object removal methods—namely, their reliance on initial 3D reconstruction, multi-view geometry priors, and inconsistent appearance modeling—this paper proposes the first reconstruction-free explicit inpainting framework. Methodologically: (1) it introduces an explicit prior alignment mechanism in pixel space to enforce cross-view geometric consistency; (2) it designs a scale-invariant depth loss to bypass scale and translation calibration inherent in monocular depth estimation; and (3) it integrates a lightweight foundation model with multi-view consistency supervision. Experiments demonstrate that our approach achieves state-of-the-art performance in both reconstruction accuracy and view consistency. Moreover, it trains three times faster than the current fastest method, significantly reducing computational overhead and deployment complexity.

0 citationsRead paper

Recent publications

Latest Papers

Kruskal-EDS: Edge Dynamic Stratification

Mar 02, 2026

This work addresses the inefficiency of the classical Kruskal’s algorithm, whose Θ(m log m) time complexity—stemming from global edge sorting—becomes prohibitive on sparse graphs or those with heavy-tailed weight distributions. The authors propose a dynamic layered variant of Kruskal’s algorithm that circumvents full sorting through a three-stage pipeline: first, sampling to estimate the weight distribution; second, adaptively constructing a layered structure inspired by the Birkhoff ergodic theorem to dynamically determine the optimal number of layers k*; and third, processing edges layer-by-layer with early termination. By integrating statistical sampling, binary search over quantile-based boundaries, and per-layer sorting, the method achieves near-linear complexity on sparse or skewed-weight graphs. Experiments across 14 graph types confirm correctness in 12 cases, demonstrating up to a 10× speedup over standard Kruskal and a 33× reduction in sorting operations.

0 citationsRead paper

Formal Semantic Control over Language Models

Jan 31, 2026

Current language models often produce semantic representations that lack interpretability and controllability, hindering localized, quasi-symbolic, and compositional semantic manipulation. To address this limitation, this work proposes a novel approach within the variational autoencoder (VAE) framework that explicitly models the geometric structure of the latent space to achieve disentangled, isolated, and directionally controllable semantic features. By systematically enhancing the interpretability and structural organization of the latent space, the method enables precise and reliable semantic control in tasks such as sentence generation and explanatory natural language inference (Explanatory NLI). This advancement significantly improves the model’s capacity for high-level semantic manipulation while maintaining fidelity to the underlying linguistic structures.

0 citationsRead paper

VEIGAR: View-consistent Explicit Inpainting and Geometry Alignment for 3D object Removal

Jun 13, 2025

To address the limitations of existing 3D object removal methods—namely, their reliance on initial 3D reconstruction, multi-view geometry priors, and inconsistent appearance modeling—this paper proposes the first reconstruction-free explicit inpainting framework. Methodologically: (1) it introduces an explicit prior alignment mechanism in pixel space to enforce cross-view geometric consistency; (2) it designs a scale-invariant depth loss to bypass scale and translation calibration inherent in monocular depth estimation; and (3) it integrates a lightweight foundation model with multi-view consistency supervision. Experiments demonstrate that our approach achieves state-of-the-art performance in both reconstruction accuracy and view consistency. Moreover, it trains three times faster than the current fastest method, significantly reducing computational overhead and deployment complexity.

0 citationsRead paper